4 papers
NeuMoSync: End-to-End Neuromodulatory Control for Plasticity and Adaptability in Continual Learning
Seyed Roozbeh Razavi Rohani, Khashayar Khajavi, Wesley Chung +2
Continual learning (CL) requires models to learn tasks sequentially, yet deep neural networks often suffer from plasticity loss and poor knowledge transfer, which can impede their…
Glass Segmentation with Fusion of Learned and General Visual Features
Risto Ojala, Tristan Ellison, Mo Chen
Glass surface segmentation from RGB images is a challenging task, since glass as a transparent material distinctly lacks visual characteristics. However, glass segmentation is crit…
Self-Supervised Masked Autoencoders with Dense-Unet for Coronary Calcium Removal in limited CT Data
Mo Chen
Coronary calcification creates blooming artifacts in Computed Tomography Angiography (CTA), severely hampering the diagnosis of lumen stenosis. While Deep Convolutional Neural Netw…
Preserving Plasticity in Continual Learning with Adaptive Linearity Injection
Seyed Roozbeh Razavi Rohani, Khashayar Khajavi, Wesley Chung +2
Loss of plasticity in deep neural networks is the gradual reduction in a model's capacity to incrementally learn and has been identified as a key obstacle to learning in non-statio…